Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
sven357
Paddle
提交
84d9c690
P
Paddle
项目概览
sven357
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
通知
1
Star
1
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
1
列表
看板
标记
里程碑
合并请求
0
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
P
Paddle
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
1
Issue
1
列表
看板
标记
里程碑
合并请求
0
合并请求
0
Pages
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
体验新版 GitCode,发现更多精彩内容 >>
提交
84d9c690
编写于
2月 12, 2018
作者:
C
chengduoZH
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
follow comments of yaming and qingqing
上级
e9fa7a7b
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
86 addition
and
130 deletion
+86
-130
python/paddle/v2/fluid/layers/detection.py
python/paddle/v2/fluid/layers/detection.py
+86
-97
python/paddle/v2/fluid/nets.py
python/paddle/v2/fluid/nets.py
+0
-33
未找到文件。
python/paddle/v2/fluid/layers/detection.py
浏览文件 @
84d9c690
...
...
@@ -151,36 +151,36 @@ def prior_box(inputs,
<https://arxiv.org/abs/1512.02325>`_ .
Args:
inputs(list): The list of input Variables, the format
inputs(list
|tuple
): The list of input Variables, the format
of all Variables is NCHW.
image(Variable): The input image data of PriorBoxOp,
the layout is NCHW.
min_ratio(int): the min ratio of generated prior boxes.
max_ratio(int): the max ratio of generated prior boxes.
aspect_ratios(list): the aspect ratios of generated prior
aspect_ratios(list
|tuple
): the aspect ratios of generated prior
boxes. The length of input and aspect_ratios must be equal.
base_size(int): the base_size is used to get min_size
and max_size according to min_ratio and max_ratio.
step_w(list
, optional, default=
None): Prior boxes step
step_w(list
|tuple|
None): Prior boxes step
across width. If step_w[i] == 0.0, the prior boxes step
across width of the inputs[i] will be automatically calculated.
step_h(list
, optional, default=
None): Prior boxes step
step_h(list
|tuple|
None): Prior boxes step
across height, If step_h[i] == 0.0, the prior boxes
step across height of the inputs[i] will be automatically calculated.
offset(float, optional, default=0.5): Prior boxes center offset.
variance(list
, optional, default=
[0.1, 0.1, 0.1, 0.1]): the variances
variance(list
|tuple|
[0.1, 0.1, 0.1, 0.1]): the variances
to be encoded in prior boxes.
flip(bool
, optional, default=
False): Whether to flip
flip(bool
|
False): Whether to flip
aspect ratios.
clip(bool, optional, default=False): Whether to clip
out-of-boundary boxes.
min_sizes(list
, optional, default=
None): If `len(inputs) <=2`,
min_sizes(list
|tuple|
None): If `len(inputs) <=2`,
min_sizes must be set up, and the length of min_sizes
should equal to the length of inputs.
max_sizes(list
, optional, default=
None): If `len(inputs) <=2`,
max_sizes(list
|tuple|
None): If `len(inputs) <=2`,
max_sizes must be set up, and the length of min_sizes
should equal to the length of inputs.
name(str
, optional,
None): Name of the prior box layer.
name(str
|
None): Name of the prior box layer.
Returns:
boxes(Variable): the output prior boxes of PriorBox.
...
...
@@ -252,7 +252,16 @@ def prior_box(inputs,
out
=
ops
.
reshape
(
x
=
input
,
shape
=
new_shape
)
return
out
assert
isinstance
(
inputs
,
list
),
'inputs should be a list.'
def
_is_list_or_tuple_
(
data
):
return
(
isinstance
(
data
,
list
)
or
isinstance
(
data
,
tuple
))
def
_is_list_or_tuple_and_equal
(
data
,
length
,
err_info
):
if
not
(
_is_list_or_tuple_
(
data
)
and
len
(
data
)
==
length
):
raise
ValueError
(
err_info
)
if
not
_is_list_or_tuple_
(
inputs
):
raise
ValueError
(
'inputs should be a list or tuple.'
)
num_layer
=
len
(
inputs
)
if
num_layer
<=
2
:
...
...
@@ -269,26 +278,25 @@ def prior_box(inputs,
max_sizes
=
[
base_size
*
.
20
]
+
max_sizes
if
aspect_ratios
:
if
not
(
isinstance
(
aspect_ratios
,
list
)
and
len
(
aspect_ratios
)
==
num_layer
):
raise
ValueError
(
'aspect_ratios should be list and the length of inputs '
'and aspect_ratios should be the same.'
)
_is_list_or_tuple_and_equal
(
aspect_ratios
,
num_layer
,
'aspect_ratios should be list and the length of inputs '
'and aspect_ratios should be the same.'
)
if
step_h
:
if
not
(
isinstance
(
step_h
,
list
)
and
len
(
step_h
)
==
num_layer
):
raise
ValueError
(
'step_h should be list and the length of inputs and '
'step_h should be the same.'
)
_is_list_or_tuple_and_equal
(
step_h
,
num_layer
,
'step_h should be list and the length of inputs and '
'step_h should be the same.'
)
if
step_w
:
if
not
(
isinstance
(
step_w
,
list
)
and
len
(
step_w
)
==
num_layer
):
raise
ValueError
(
'step_w should be list and the length of inputs and '
'step_w should be the same.'
)
_is_list_or_tuple_and_equal
(
step_w
,
num_layer
,
'step_w should be list and the length of inputs and '
'step_w should be the same.'
)
if
steps
:
if
not
(
isinstance
(
steps
,
list
)
and
len
(
steps
)
==
num_layer
):
raise
ValueError
(
'steps should be list and the length of inputs and '
'step_w should be the same.'
)
_is_list_or_tuple_and_equal
(
steps
,
num_layer
,
'steps should be list and the length of inputs and '
'step_w should be the same.'
)
step_w
=
steps
step_h
=
steps
...
...
@@ -298,13 +306,13 @@ def prior_box(inputs,
min_size
=
min_sizes
[
i
]
max_size
=
max_sizes
[
i
]
aspect_ratio
=
[]
if
not
isinstance
(
min_size
,
list
):
if
not
_is_list_or_tuple_
(
min_size
):
min_size
=
[
min_size
]
if
not
isinstance
(
max_size
,
list
):
if
not
_is_list_or_tuple_
(
max_size
):
max_size
=
[
max_size
]
if
aspect_ratios
:
aspect_ratio
=
aspect_ratios
[
i
]
if
not
isinstance
(
aspect_ratio
,
list
):
if
not
_is_list_or_tuple_
(
aspect_ratio
):
aspect_ratio
=
[
aspect_ratio
]
box
,
var
=
_prior_box_
(
input
,
image
,
min_size
,
max_size
,
aspect_ratio
,
...
...
@@ -354,26 +362,26 @@ def multi_box_head(inputs,
MultiBox Detector)<https://arxiv.org/abs/1512.02325>`_ .
Args:
inputs(list): The list of input Variables, the format
inputs(list
|tuple
): The list of input Variables, the format
of all Variables is NCHW.
num_classes(int): The number of c
als
s.
min_sizes(list
, optional, default=None): The length
of
min_size
is used to compute the the number of prior
box.
num_classes(int): The number of c
lasse
s.
min_sizes(list
|tuple|None): The number
of
min_size
s is used to compute the number of predicted
box.
If the min_size is None, it will be computed according
to min_ratio and max_ratio.
max_sizes(list
, optional, default=None): The length of max_size
is used to compute the the number of pr
ior
box.
min_ratio(int
): If the min_sizes is None, min_ratio and min
_ratio
max_sizes(list
|tuple|None): The number of max_sizes
is used to compute the the number of pr
edicted
box.
min_ratio(int
|None): If the min_sizes is None, min_ratio and max
_ratio
will be used to compute the min_sizes and max_sizes.
max_ratio(int
): If the min_sizes is None, min
_ratio and min_ratio
max_ratio(int
|None): If the min_sizes is None, max
_ratio and min_ratio
will be used to compute the min_sizes and max_sizes.
aspect_ratios(list): The number of the aspect ratios is used to
aspect_ratios(list
|tuple
): The number of the aspect ratios is used to
compute the number of prior box.
base_size(int): the base_size is used to get min_size
and max_size according to min_ratio and max_ratio.
flip(bool
, optional, default=
False): Whether to flip
flip(bool
|
False): Whether to flip
aspect ratios.
name(str
, optional,
None): Name of the prior box layer.
name(str
|
None): Name of the prior box layer.
Returns:
...
...
@@ -397,52 +405,33 @@ def multi_box_head(inputs,
flip=True)
"""
def
_conv_with_bn_
(
input
,
conv_num_filter
,
conv_padding
=
1
,
conv_filter_size
=
3
,
conv_stride
=
1
,
conv_act
=
None
,
param_attr
=
None
,
conv_with_batchnorm
=
False
,
conv_batchnorm_drop_rate
=
0.0
,
use_cudnn
=
True
):
conv2d
=
nn
.
conv2d
(
input
=
input
,
num_filters
=
conv_num_filter
,
filter_size
=
conv_filter_size
,
padding
=
conv_padding
,
stride
=
conv_stride
,
param_attr
=
param_attr
,
act
=
conv_act
,
use_cudnn
=
use_cudnn
)
if
conv_with_batchnorm
:
conv2d
=
nn
.
batch_norm
(
input
=
conv2d
)
drop_rate
=
conv_batchnorm_drop_rate
if
abs
(
drop_rate
)
>
1e-5
:
conv2d
=
nn
.
dropout
(
x
=
conv2d
,
dropout_prob
=
drop_rate
)
def
_is_equal_
(
len1
,
len2
,
err_info
):
if
not
(
len1
==
len2
):
raise
ValueError
(
err_info
)
return
conv2d
def
_is_list_or_tuple_
(
data
):
return
(
isinstance
(
data
,
list
)
or
isinstance
(
data
,
tuple
))
if
not
(
isinstance
(
inputs
,
list
)
):
raise
ValueError
(
'inputs should be a list.'
)
if
not
_is_list_or_tuple_
(
inputs
):
raise
ValueError
(
'inputs should be a list
or tuple
.'
)
if
min_sizes
is
not
None
:
if
not
(
len
(
inputs
)
==
len
(
min_sizes
)):
raise
ValueError
(
'the length of min_sizes '
'and inputs should be the same.'
)
_is_equal_
(
len
(
inputs
),
len
(
min_sizes
),
'the length of min_sizes '
'and inputs should be equal.'
)
if
max_sizes
is
not
None
:
if
not
(
len
(
inputs
)
==
len
(
max_sizes
)):
raise
ValueError
(
'the length of max_sizes '
'and inputs should be the same.'
)
_is_equal_
(
len
(
inputs
),
len
(
max_sizes
),
'the length of max_sizes '
'and inputs should be equal.'
)
if
aspect_ratios
is
not
None
:
if
not
(
len
(
inputs
)
==
len
(
aspect_ratios
)):
raise
ValueError
(
'the length of aspect_ratios '
'and inputs should be the same.'
)
_is_equal_
(
len
(
inputs
),
len
(
aspect_ratios
),
'the length of aspect_ratios '
'and inputs should be equal.'
)
if
min_sizes
is
None
:
# If min_sizes is None, min_sizes and max_sizes
...
...
@@ -464,22 +453,23 @@ def multi_box_head(inputs,
mbox_confs
=
[]
for
i
,
input
in
enumerate
(
inputs
):
min_size
=
min_sizes
[
i
]
if
type
(
min_size
)
is
not
list
:
if
not
_is_list_or_tuple_
(
min_size
)
:
min_size
=
[
min_size
]
max_size
=
[]
if
max_sizes
is
not
None
:
max_size
=
max_sizes
[
i
]
if
type
(
max_size
)
is
not
list
:
if
not
_is_list_or_tuple_
(
max_size
)
:
max_size
=
[
max_size
]
if
not
(
len
(
max_size
)
==
len
(
min_size
)):
raise
ValueError
(
'max_size and min_size should have same length.'
)
_is_equal_
(
len
(
max_size
),
len
(
min_size
),
'the length of max_size and min_size should be equal.'
)
aspect_ratio
=
[]
if
aspect_ratios
is
not
None
:
aspect_ratio
=
aspect_ratios
[
i
]
if
type
(
aspect_ratio
)
is
not
list
:
if
not
_is_list_or_tuple_
(
aspect_ratio
)
:
aspect_ratio
=
[
aspect_ratio
]
# get the number of prior box on each location
...
...
@@ -499,25 +489,24 @@ def multi_box_head(inputs,
if
share_location
:
num_loc_output
*=
num_classes
mbox_loc
=
_conv_with_bn_
(
mbox_loc
=
nn
.
conv2d
(
input
=
input
,
conv_num_filter
=
num_loc_output
,
conv_padding
=
pad
,
conv_stride
=
stride
,
conv_filter_size
=
kernel_size
,
conv_with_batchnorm
=
use_batchnorm
)
num_filters
=
num_loc_output
,
filter_size
=
kernel_size
,
padding
=
pad
,
stride
=
stride
)
mbox_loc
=
nn
.
transpose
(
mbox_loc
,
perm
=
[
0
,
2
,
3
,
1
])
mbox_locs
.
append
(
mbox_loc
)
# get conf_loc
num_conf_output
=
num_priors_per_location
*
num_classes
conf_loc
=
_conv_with_bn_
(
conf_loc
=
nn
.
conv2d
(
input
=
input
,
conv_num_filter
=
num_conf_output
,
conv_padding
=
pad
,
conv_stride
=
stride
,
conv_filter_size
=
kernel_size
,
conv_with_batchnorm
=
use_batchnorm
)
num_filters
=
num_conf_output
,
filter_size
=
kernel_size
,
padding
=
pad
,
stride
=
stride
)
conf_loc
=
nn
.
transpose
(
conf_loc
,
perm
=
[
0
,
2
,
3
,
1
])
mbox_confs
.
append
(
conf_loc
)
...
...
python/paddle/v2/fluid/nets.py
浏览文件 @
84d9c690
...
...
@@ -18,7 +18,6 @@ __all__ = [
"sequence_conv_pool"
,
"glu"
,
"scaled_dot_product_attention"
,
"img_conv_with_bn"
,
]
...
...
@@ -108,38 +107,6 @@ def img_conv_group(input,
return
pool_out
def
img_conv_with_bn
(
input
,
conv_num_filter
,
conv_padding
=
1
,
conv_filter_size
=
3
,
conv_stride
=
1
,
conv_act
=
None
,
param_attr
=
None
,
conv_with_batchnorm
=
False
,
conv_batchnorm_drop_rate
=
0.0
,
use_cudnn
=
True
):
"""
Image Convolution Group, Used for vgg net.
"""
conv2d
=
layers
.
conv2d
(
input
=
input
,
num_filters
=
conv_num_filter
,
filter_size
=
conv_filter_size
,
padding
=
conv_padding
,
stride
=
conv_stride
,
param_attr
=
param_attr
,
act
=
conv_act
,
use_cudnn
=
use_cudnn
)
if
conv_with_batchnorm
:
conv2d
=
layers
.
batch_norm
(
input
=
conv2d
)
drop_rate
=
conv_batchnorm_drop_rate
if
abs
(
drop_rate
)
>
1e-5
:
conv2d
=
layers
.
dropout
(
x
=
conv2d
,
dropout_prob
=
drop_rate
)
return
conv2d
def
sequence_conv_pool
(
input
,
num_filters
,
filter_size
,
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录